Analysis of litter decomposition in an alpinetundra

David M. Bryant, Elisabeth A. Holland, Timothy R. Seastedt,and Marilyn D. Walker

Abstract: Decomposition of plant litter regulates nutrient cycling and transfers of fixed carbon to soil organic matterpools in terrestrial ecosystems. Climate, as well as factors of intrinsic litter chemistry, often govern the rate ofdecomposition and thus the dynamics of these processes. Initial concentrations of nitrogen and recalcitrant carboncompounds in plant litter are good predictors of litter decomposition rates in many systems. The effect of exogenousnitrogen availability on decay rates, however, is not well defined. Microclimate factors vary widely within alpinetundra sites, potentially affecting litter decay rates at the local scale. A controlled factorial experiment was performedto assess the influence of landscape position and exogenous nitrogen additions on decomposition of surface foliage andburied root litter in an alpine tundra in the Front Range of the Rocky Mountains in Colorado, U.S.A. Litter bags wereplaced in three communities representing a gradient of soil moisture and temperature. Ammonium nitrate was appliedonce every 30 days at a rate of 20 g N·m–2 during the 3-month growing season. Data, as part of the Long-TermInter-site Decomposition Experiment Team project, were analyzed to ascertain the effects of intrinsic nitrogen andcarbon fraction chemistry on litter decay in alpine systems. Soil moisture was found to be the primary controllingfactor in surface litter mass loss. Root litter did not show significant mass loss following first growing season.Nitrogen additions had no effect on nitrogen retention, or decomposition, of surface or buried root litter compared withcontrols. The acid-insoluble carbon fraction was a good predictor of mass loss in surface litters, showing a strongnegative correlation. Curiously, N concentration appeared to retard root decomposition, although degrees of freedomlimit the confidence of this observation. Given the slow rate of decay and N loss from root litter, root biomass appearsto be a long-term reservoir for C and N in the alpine tundra.

Key words: litter decomposition, alpine tundra, nitrogen deposition, LIDET, Niwot Ridge.1304

Résumé: Dans les écosystèmes terrestres, la décomposition des litières végétales règle le cyclage des nutriments et lestranferts du carbone fixé aux réserves de matière organique du sol. Le climat, ainsi que des facteurs intrinsèques à lachimie de la litière, déterminent le taux de décomposition et ainsi la dynamique de ces processus. Dans plusieurssystèmes, la teneur initiale en azote et les composés carbonés récalcitrants sont de bons éléments de prédiction destaux de décomposition. L’effet d’une disponibilité d’azote exogène sur les taux de décomposition est cependant maldéfini. Les facteurs microclimatiques varient beaucoup sur les sites de tundra alpine, ce qui est susceptible d’affecterles taux de décomposition de la litière à l’échelle locale. Les auteurs ont conduit une expérience factorielle contrôléepour évaluer l’influence de la position dans le paysage et d’additions d’azote sur la décomposition en surface defeuillage et en profondeur de litière racinaire, dans une tundra alpine située dans le Front Range des montagnesRocheuses au Colorado, U.S.A. Les sacs de litières ont été placés dans trois communautés représentant un gradientd’humidité et de température du sol. Du nitrate d’ammonium a été appliqué à tous les 30 jours à raison de 20 g N·m–2

au cours de la saison de croissance de 3 mois. Dans le cadre du projet en équipe « Long-Term Inter-siteDecomposition », les auteurs ont analysé les données pour préciser les effets de l’azote intrinsèque et de la chimie dela fraction carbonée sur la décomposition des litières en milieu alpin. Il apparait que l’humidité du sol est le premierfacteur contrôlant la perte de masse de la litière en surface du sol. La litière racinaire ne montre pas de perte de massesignificative après la première saison. Les additions d’azote n’ont pas d’effet sur la rétention de l’azote, ou ladécomposition, de la litière de surface ou racinaire enfouie, comparativement aux témoins. La fraction carbonée solubleà l’acide est un bon critère de prédiction de la perte de masse pour la litière de surface, montrant une forte corrélationnégative. Curieusement, la concentration en N semble retarder la décomposition des racines, bien que les degrés deliberté limitent la confiance en cette observation. Étant donné le faible taux de décomposition et de la perte en N deslitières racinaires, la biomasse racinaire semble constituer une réserve à long terme de C et de N en milieu alpin.

Can. J. Bot.76: 1295–1304 (1998) © 1998 NRC Canada

1295

Received October 1, 1997.

D.M. Bryant, 1 T.R. Seastedt, and M.D. Walker.Institute of Arctic and Alpine Research, University of Colorado, Boulder,CO 80309, U.S.A.E.A. Holland. National Center for Atmospheric Research, Boulder, CO 80307, U.S.A.

1Author to whom all correspondence should be addressed. Present address: Department of Natural Resources, 215 James Hall,University of New Hampshire, Durham, NH 03824, U.S.A. e-mail: dmbryant@christa.unh.edu

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

Mots clés: décomposition des litières, tundra alpine, addition d’azote, LIDET, Niwot Ridge.

[Traduit par la Rédaction]

Bryant et al.

Decomposition of plant litter is an important regulator ofnutrient cycling and carbon (C) flux in terrestrial ecosystems(Swift et al. 1979; Aber and Melillo 1991; Schlesinger1991). The rate of decomposition often correlates with theinitial relative concentrations of C compounds and the nitro-gen (N) content of plant litter (Fogel and Cromack 1977;Aber and Melillo 1982; Melillo et al. 1982, 1989; Melilloand Aber 1984; Berendse et al. 1987; Aber et al. 1990; Tay-lor et al. 1989). Thus, indices of carbon to nitrogen ratio(C/N) (Fogel and Cromack 1977; Berendse et al. 1987; Tay-lor et al. 1989), lignin to nitrogen ratio (L/N) (Aber andMelillo 1982; Aber et al. 1990) and lignocellulose index(lignin/lignin + cellulose) (Melillo et al. 1989) have beenused to predict decay rates in temperate forest and grasslandsystems.

The C/N ratio has also been shown to influence N dynam-ics of decaying litter. When litter C/N is high, decomposermicrobes immobilize N and thus may delay N release (Aberand Melillo 1982; Melillo and Aber 1984). Low C/N ratiosrelieve microbial N limitation, potentially accelerating de-cay, N release, and increasing N availability (Aber et al.1990).

While the relationship between the N content of litter anddecay rate has been observed in numerous studies, no clearconnection has been found between decomposition rate andexogenous N availability (Fog 1988). Comparisons at theplot level have occasionally produced positive correlationsof N availability with litter decomposition rate (Kelly andHenderson 1978; Hunt et al. 1988). When data were com-pared across sites, however, the strengths of the relationshipswere reduced (McClaugherty et al. 1985; Hunt et al. 1988).

Meentmeyer (1978) observed that the negative effect oflignin content on decay rate increases with actual evapo-transpiration (AET) across a latitudinal gradient. Thus, vari-ations in environment appear to interact with litter chemistryto regulate decomposition rates.

The Niwot Ridge tundra ecosystem in the Colorado FrontRange consists of a mosaic of plant communities, the distri-bution of which is influenced primarily by snow depth(Walker et al. 1993). Variation in topography, snow depth,soil moisture, and plant biomass combine to create a rangeof microclimates across the alpine landscape. In addition, at-mospheric N deposition has been increasing at this site andmay exceed 30 times preindustrial levels (Lewis et al. 1984;Bowman 1992; Sievering et al. 1992). Consequently, the po-tential exists for significant variation in decomposition andN cycling rates among communities.

The influence of intrinsic chemistry and environmentalvariables on litter decomposition have been extensively in-vestigated in temperate forests (e.g., McClaugherty and Berg1987), boreal forests (e.g., van Cleve 1974), grasslands (e.g.,Seastedt et al. 1992), agroecosystems (e.g., Holland andColeman 1987), and arctic tundra ecosystems (e.g., Hobbie1996). The current literature on decomposition in alpine tun-dra, however, is sparse (e.g., O’Lear and Seastedt 1994).

The objectives of the current research were to (i) deter-mine the variation in decomposition rates over the range oftemperature and moisture found at the Niwot Ridge site;(ii ) test the effect of exogenous N additions on decomposi-tion of native litter; and (iii ) determine the influence of litterchemistry on decay rate and N dynamics. These questionsare addressed separately for surface foliage and buried fineroot litter.

Site descriptionNiwot Ridge is a UNESCO Biosphere Reserve and a Long-Term

Ecological Research Site (LTER). The site is located at 3510 m el-evation, 5 km east of the Continental Divide, and 8 km west ofWard, Colo. (E 499.700, N 4,433,900 UTM). This study was per-formed in a 25-ha saddle area between two knolls, thus providing atoposequence of varying slope and aspect. Average annual precipi-tation is 930 mm, of which 82% falls as snow between late Sep-tember and early June. Snow depth is variable across the saddleresulting from redistribution by high winds. Mean annual air tem-perature measured at a climate station 300 m above the saddle siteis –3.8°C. The frost-free growing season averages 47 days begin-ning in late June and extending to mid-August (Walker et al.1994).

Of the six major vegetation types described by May and Webber(1975) at the Saddle site, the dry, mesic, and wet communities oc-cupy the largest area. Dry meadows are located on the wind-sweptwest facing knoll and are dominated by the sedge,Kobresiamyosuroides, and the forb,Acomostylis rossii. The mesic meadowcommunities occupy the center portions of the saddle whereAcomostylis rossiiand a grass,Deschampsia caespitosa,are themost abundant species. The wet meadows are found adjacent to thesnow fields at the base of the west knoll and are dominated by theforb, Caltha leptasepala,and a sedge,Carex scopulorum. Thesecommunities constitute a complex environmental gradient of mois-ture, plant biomass, snow depth, and topography among other vari-ables.

Experimental design

Nitrogen × landscape treatmentA factorial design was implemented to measure the effects of

landscape position and N additions on decomposition of native sur-face foliage and buried fine root litter. Six 1-m2 plots were locatedrandomly in each of the three community types (dry, mesic, andwet) resulting in a total of 18 plots. Nitrogen treatment bags wereplaced 75 cm downslope from the control bags in a split-plot ar-rangement to prevent runoff contamination of controls. Five of the10 surface litterbags at each plot received N treatments, and fivewere exposed to ambient N levels. Four plots in each communitytype received eight root litter bags, four for each level of N addi-tion.

Chemistry treatmentThe influence of intrinsic chemistry on litter decay was deter-

mined using data obtained from the Long-Term Intersite Decompo-sition Experiment Team (LIDET) project. The study was initiatedin 1990 as a large-scale, cross-site investigation, focusing on theinfluence of climate and litter chemistry on decay rates.Twenty-nine species of litter were incubated at 28 sites represent-

© 1998 NRC Canada

1296 Can. J. Bot. Vol. 76, 1998

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

ing a latitudinal gradient from Central America to the North Slopeof Alaska and elevational gradient from sea level to 3510 m.

Litter at the Niwot site was incubated within the mesic meadowcommunity at four replicate plots. Baseline chemistry data wereanalyzed from pooled subsamples of each species and litter typefor proximate C fraction (PCF) and N. The LIDET team chose sixspecies of foliar litter (Acer saccharumMarsh.,Drypetes glauca,Pinus resinosaAit., Quercus prinusL., Thuja plicataDonn., andTriticum aestivumL.) and three species of root litter (Drypetesglauca, Pinus elliotii Engelm., andAndropogon gerardi) for therepeated annual collections.

Field methods

Native litterFoliar litter consisted of a composite of graminoid litters from

three sources. The major constituent was obtained from a pool ofdried, archived material from live peak biomass collections occur-ring from 1984 through 1992. Methods of collection are describedin Walker et al. (1994). Senesced litter was collected from the fieldin spring 1993 and combined with senescedDeschampsialitterfrom greenhouse grown plants. Senesced and green litters wereplaced in separate 40-L polyethylene bags. All litter was brought to10% gravimetric water content by lightly spraying with deionizedwater to minimize fragmentation during weighing and handling.The litter was stored at 4°C for 24 h to allow for equilibration.Green and senesced foliage was combined at 1.5:1 ratio into 2-gsamples and placed in 10 × 20 cm fiberglass screen litter bags of1 mm mesh size. Litter bags were stapled closed and dried at 60°Cto constant weight. Initial litter weight was obtained by subtractingthe bag weight (including staples) from total weight followingoven drying. An additional 30 foliage litter bags were prepared us-ing knit polyester cloth with a mesh of <0.3 mm. These bags wereplaced at six of the paired plots as a control on litter bag mesh size.Surface foliage bags were secured to the tundra surface with nails.

Roots were collected by coring at Niwot Ridge and were notidentified to species. Fine roots (<2 mm diameter) were separatedfrom coarse roots (>2 mm diameter), but no attempt was made toseparate live and dead roots. The remaining fine roots were hy-drated as above and homogenized by hand mixing. A 2-g samplewas placed in each 15 × 20 cm knit polyester litter bag (mesh <0.3 mm) and oven-dried to constant weight at 60°C. A wedge (3 ×15 × 15 cm) was cut and removed from the soil, and the litter bagwas inserted vertically. The wedge was replaced adjacent to the lit-ter bag and pressed into place. The aluminum identification tagwas left exposed at the surface to locate the bag for retrieval.

The treated bags of each plot received applications of N fertil-izer in the form of aqueous NH4NO3. Treatments of 20 g N·m–2

were applied in June, July, and August of each year for a total of60 g N·m–2·year–1. Surface litter bags were collected five times, attime of placement, fall 1993, spring 1994, fall 1994, and fall 1995.Time-zero (t0) samples allowed correction of losses incurred dur-ing transport and weighing, as well as measurement of baselinechemical concentrations. Buried root litter bags were handled iden-tically to surface bags with the exception of a time-zero harvest.Initial chemistry for fine roots was measured from subsamples offine root litter taken while constructing root litter bags.

Soil temperature measurements were taken at a subset of fiveplots that encompass the range of slope and aspect for the Saddleresearch area. Thermisters were placed at 2- and 10-cm depths.Automated measurements were made starting June 16, 1994,through September 8, 1994, and stored in a Datapod 212 data log-ger (Omnidata International, Inc. Logan, Utah). Daily averages arereported here.

Soil moisture measurements were taken weekly at all litter bagsites during the 1994 growing season using the time-domain

reflectometry (TDR) method described in Taylor and Seastedt(1994).

LIDET litterFoliage was collected as senesced or abscised leaves and

air-dried with the exception ofDrypetes glauca, which wasoven-dried. All material was shipped to the Forest Research Labo-ratory at Oregon State University (OSU), headquarters of theLIDET project. Roots were collected by excavation except in thecase ofAndropogon gerardii, which were collected from materialexposed along stream banks. Surface litter bags were 20 × 20 cm2

with 1-mm mesh on the top and 55-µm mm mesh on the bottom.Each surface bag received 10 g of litter and was labeled with alu-minum number tags. Bags were threaded onto cords in random or-der. Strings of litter bags were staked to the ground at four sites inthe mesic alpine meadow. Root litter bags were constructed of55-µm mesh and filled with 5–7 g of fine root litter. A vertical cutin the soil was made with a shovel and a litter bag inserted verti-cally. A second parallel cut was made, adjacent to the first, to pressthe soil firmly against the litter bag.

One litter bag of each species was collected annually from eachreplicate site. Buried root bags were washed with deionized (DI)water to remove adherent soil from the surface of the bag. Litterwas removed from bags, placed in paper bags, and oven-dried at55°C to constant weight. Samples were sent to OSU for weighingand chemical analyses. Ash content of litter was determined byloss on ignition after collection.

Laboratory methodsSamples from each harvest were returned to the lab on the same

day of collection. Any new roots were removed, samples wereoven-dried at 60°C to constant weight, and ground in a Wiley millto pass a 40-mesh screen. Native and LIDET litter samples weretreated similarly with the exception that LIDET samples wereshipped to OSU following oven drying and weighing. Subsamplesof t0 samples were composited and analyzed via the PCF method(Newman et al. 1995) to determine relative concentrations of ex-tractive, acid-soluble (ACS) and acid-insoluble (AIS) fractions.These last two fractions have previously been termed holocelluloseand lignin. While the acid-soluble fraction has been quantitativelyidentified as holocellulose (Effland 1977), lignin is a less well de-fined substance and thus is referred to here as an AIS fraction. Ashcontents were determined by the loss on ignition method. Nitrogenand C concentrations of the native litter were measured by a CarloErba CHN elemental analyzer (Fison Instruments, Danvers, Mass.)as were the initial values for the LIDET litter. Litter chemistry dataon the time course of decayed LIDET samples were not availablefrom the LIDET team as of this writing; therefore, data presentedhere on those substrates were acquired through micro-Kjeldahl di-gestion, followed by colorimetric analysis, performed by at theUniversity of New Hampshire, Durham, N.H.

Subsamples of native litter from surface and buried control plotsand dry and wet surface N treatment plots were sent to the plantpathology laboratory at OSU for analysis of fungal and bacterialbiomass. The techniques for microbial biomass determination aredescribed in Ingham and Klein (1984).

Statistical methodsData were analyzed using the JMP version 3.0 statistical soft-

ware (SAS Institute Inc., Cary, N.C.). Mean values of each vari-able presented were calculated by lumping replicates within eachcommunity type and treatment separately. Absolute differences indecay and N release were compared using the ash-free percentoriginal mass (POM) remaining at the time of harvest. The arcsinetransformation was applied to percentage data to achieve independ-ence of the mean and variance in each data set prior to statistical

© 1998 NRC Canada

Bryant et al. 1297

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

analysis. Differences in the mean absolute mass loss and mean per-cent of original N remaining (dependent variables) among commu-nity and N treatment effects were determined via model II,two-way analysis of variance (ANOVA) after Sokal and Rohlf(1995). Specific differences among community types were testedpairwise using Tukey’s honestly significant difference (HSD). Dif-ferences among LIDET species effects were tested by repeatedmeasures MANOVA and the pairwise comparison with Tukey’sHSD.

Mass loss was compared with environmental variables of soilmoisture, soil surface temperature (surface litter samples), soil

temperature at 10 cm (buried fine root litter), and snow cover byleast squares multiple regression.

Litter chemistryResults of initial PCF and N content analysis of all litters

are shown in Table 1.

Surface litter mass lossThe pattern of decay in each community produced good

fits with the exponential model for all treatments (Table 2).Percent original surface litter mass remaining in coarse-mesh litter bags varied significantly among communities forall harvests (Fig. 1A). Mass loss clearly decreases across thelandscape from wet to dry sites. Soil moisture mea-surements were significantly correlated with POM remain-ing at 12 months (p < 0.01) (Fig. 2), while mean dailysurface temperature was not.

Litter mass remaining in fine-mesh litter bags showedsimilar patterns of landscape and decomposition rates withthe coarse mesh bags early in the study. However, in the fall1994 harvest (15 months) the larger (1-mm) mesh bagsshowed a significant decrease in mass over fine-mesh bags(p < 0.05; data not shown). Landscape position is also sig-nificant in both treatments (p < 0.01) although landscape ×mesh size interaction is not.

Additions of N produced no significant differences in sur-face mass loss among communities during the course ofthese measurements (p = 0.06) (Table 2).

Foliage litter showed significant differences in massremaining among the LIDET species at each of the annualharvests for the first 3 years of the study (Fig. 1C). Differ-ences in relative decay among all species are also significant(Table 3) with the exceptionDrypetes glaucaand Acersaccharum. The pattern of decay shown in Fig. 1C appearsto follow AIS concentrations; indeed, initial AIS correlateswell with k for surface foliage (Fig. 3A). The species

© 1998 NRC Canada

1298 Can. J. Bot. Vol. 76, 1998

Species % N % Extractive % ACS % AIS

FoliageNative 1.2 22 60 18

LIDETChestnut oak (Quercus prinus) 1.0 41.0 42.0 17.0Drypetes (Drypetes glauca) 1.7 46.1 42.6 9.3Red pine (Pinus resinosa) 0.7 41.0 38.3 20.1Sugar maple (Acer saccharum) 1.1 58.7 33.5 7.8Western red cedar (Thuja plicata) 0.9 42.0 41.0 17.8Wheat (Triticum aestivum) 0.3 30.5 52.1 17.4

Fine rootsNative 1.1 15.0 50.0 35.0

LIDETBig blue-stem (Andropogon gerardi) 0.6 39.2 49.0 11.8Drypetes (Drypetes glauca) 0.9 30.1 45.3 23.9Slash pine (Pinus elliotii) 1.0 48.7 37.4 13.9

Note: See text for composition of native foliage. Extractives, polar + non-polar soluble fraction; ACS, acid soluble fraction; AIS,acid insoluble fraction.

Table 1. Initial nitrogen and proximate carbon fraction chemistry of native alpine and LIDET leaf and fine rootlitters used in litter decay experiment Niwot Ridge, Colorado, U.S.A.

k factor R2 Error p

Foliage litter

ControlsDry 0.021 0.8991 0.071 <0.0001Mesic 0.025 0.7779 0.131 <0.0001Wet 0.027 0.8655 0.103 <0.0001

N treatmentDry 0.013 0.9446 0.044 <0.0001Mesic 0.022 0.8240 0.104 <0.0001Wet 0.025 0.8293 0.114 <0.0001

Fine root litter

ControlsDry 0.0062 0.3315 0.0863 <0.01Mesic 0.0106 0.7404 0.0637 <0.0001Wet 0.0067 0.2317 0.105 <0.02

N treatmentDry 0.010 0.5663 0.083 <0.01Mesic 0.006 0.4271 0.0066 <0.0001Wet 0.004 0.003 0.148 <0.32

Note: Error terms are root mean square error of the regression.

Table 2. Exponential decay rates (k factor) of surface foliageand buried fine root litter (N treated and controls) duringincubation in three alpine communities, Niwot Ridge, Colorado,U.S.A.

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

Quercus prinus, Thuja plicata, and Triticum aestivumhavealmost identical AIS fractions and thus group closely in thisregression. Considerable variation along they axis is appar-ent suggesting other sources, but no significant correlationwith any other PCF was found. Initial N content produced apositive relationship withk but was only marginally signifi-cant (p = 0.06).

Fine root litter mass lossFollowing an initial decline during the 1993 growing sea-

son, no significant changes in mass loss of native fine rootlitter were observed over the next four harvests. No signifi-cant differences in fine root POM remaining with landscapeposition (Fig. 1B) or N treatment (Table 2) was found within

any harvest. Surface litter decay differed significantly fromfine root litter in all treatments (p < 0.0001).

Decay of fine root litter in the LIDET experiment(Fig. 1D) also produced a good fit to the exponential decayfunction (Table 2). Regression of initial chemistry againstkfactor produced no significant relationship with any PCFvalue. Initial analysis of N concentration versus rootk factorproduced a surprisingly strong negative correlation (R2 =0.9778). This relationship suffered from degrees of freedomhaving only three species in the data set. The regression wasrerun using the estimated annualk factor of each replicate at

© 1998 NRC Canada

Bryant et al. 1299

Fig. 1. Litter mass loss for surface and buried fine-root litter in an alpine landscape. Native foliage (A) and buried fine-root (B) litterswere incubated in three community types. LIDET litter (C, foliage; D, buried fine roots) were incubated in the mesic meadowcommunity. Error bars are SE.

Fig. 2. Mass loss of native alpine surface foliage litter versusgravimetric soil moisture incubated in three alpine communities.

Species k factor R2 Error p

FoliageP. resinosa 0.039 0.6851 0.031 <0.0001T. plicata 0.067 0.8689 0.030 <0.0001T. aestivum 0.093 0.8663 0.044 <0.0001Q. prinus 0.116 0.8096 0.067 <0.0001A. saccharum 0.157 0.7336 0.114 <0.0001D. glauca 0.214 0.7358 0.153 <0.0001

Fine rootsP. elliotii 0.059 0.4807 0.073 <0.01D. glauca 0.124 0.6420 0.110 <0.01A. gerardii 0.193 0.5781 0.197 <0.001

Note: Error terms are root mean square error of the regression,p valuerefers to significance fit to regression model.

Table 3. Exponential decay rates of LIDET litter following3-year incubation in a mesic alpine meadow, Niwot Ridge,Colorado, U.S.A.

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

year 3 (k = ln(POM)/t). The R2 of the regression sufferedslightly (see Fig. 5) but was still highly significant (p <0.001).

Nitrogen dynamics

Surface litterRelease of N from decaying litter occurred in surface con-

trol plots during the first growing season. Figure 4A showsthe percent of original N (PON) at each harvest. The patternof N release was similar to that of mass loss across the land-scape. Therefore, differences in N content were highly sig-nificant among communities (p < 0.001). No significantdifference in PON remaining was observed during either the12- or 15-month harvests, as surface litter in all communitiesconverged to a mean of 61.5 ± 8.36% (mean ± SD) originalN. Over the next 12 months, N contents diverged as littersaccumulated N at different rates. However, differencesamong landscape positions or treatments were not signifi-cant. This observation compares well with results of micro-bial biomass measurements on control litter bags at12 months (Table 4). Microbial biomass (fungal + bacterial)did not correlate with landscape position, litter mass remain-ing, N treatments, or N dynamics at 15 months.

The six species of LIDET surface litter also showed lossof N during the first 12 months (Fig. 4C). Four of the sixsurface litters began accumulating N by the following yearwith the exception ofDrypetes glaucaand P. resinosa,which did not show N accumulation until year 3. WhilePON is a good illustrator of N dynamics relative to initialcontent, the absolute quantity of N lost or gained is a better

indicator of the N immobilization potential among species.Furthermore, recent studies with15N tracers show thatchanges in N content are the net result of simultaneous gainsand losses of N by decaying litter (Berg 1987; van Vuurenand van Der Veerden 1992). Therefore, regressions of abso-lute net N immobilization were performed against speciesPCF chemistry and initial N content. Initial N contentshowed a negative correlation with N immobilization fol-lowing the first year of incubation, but was not significant.Content of extractives however showed a highly significantnegative (p < 0.0001) correlation to net N immobilizationduring this period (Fig. 6). The regression against the ACSfraction was likewise significant, but the slope was positive,suggesting merely a reflection of the correlation of ACSwith extractives. This relationship declined in significance inthe following two harvests.

Fine root litterNative buried fine root litter released N in the first two

growing seasons (Fig. 6B). Regardless of landscape positionor N treatment, root litter retained an average of 79.3 ±22.53% original N mass. Native root litter began accumulat-ing N in all communities following 15 months incubation.No significant difference in N content was found amonglandscape positions. As in surface litter, mean values of C/Nfor root litter did not vary significantly from the originalmean value of 41.

Fungal and bacterial biomass were significantly higher inburied roots than in surface litter. In addition, fungal to bac-terial biomass ratios were two orders of magnitude higher infine roots than in surface litter (Table 4). As with surfacelitter though, no correlations were found with landscape po-sition or N treatment.

Following the same pattern as surface litter, LIDET rootspecies lost N during the first year (Fig. 4D); nitrogencontent was stable during the second year; and all spe-cies showed N accumulation by year 3. Nitrogen accumula-tion in Drypetes glaucaroots was significantly greaterthan P. elliotii roots although neither species differedfrom Andropogon gerardii. Net N immobilization in fineroot litter was not significantly correlated with any PCFfraction.

© 1998 NRC Canada

1300 Can. J. Bot. Vol. 76, 1998

Fig. 3. Relationships of exponential slope constant (k factor)and initial acid insoluble (AIS) fraction (A) and initial [N](B) of LIDET surface litters. Error bars are SE.

Fungus Bacteria Ratio

N Mean SE Mean SE Mean SE

FoliageDry 5 11 0.7 162 23.0 0.08 0.01Mesic 5 13 1.5 244 46.6 0.06 0.01Wet 5 20 7.4 290 63.6 0.08 0.04

Fine rootsDry 3 418 60.7 44 2.4 9.47 1.10Mesic 2 390 31.3 67 9.3 5.96 0.36Wet 4 268 24.5 65 9.9 4.95 1.46

Table 4. Microbial biomass (mg·g–1 dry wt.) for surface foliageand buried fine root litter in three alpine communities, NiwotRidge, Colorado, U.S.A.

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

The differences in litter decay rates observed across com-munity types are apparently driven by variation in soil mois-ture (Figs. 1A, 1B, and 2). In other systems, Vitousek et al.(1994) found temperature to have a greater influence thanmoisture on litter decomposition across an elevational gradi-ent on Mauna Loa, Hawaii. Witkamp (1966) found similarresults in temperate forests. Both observations agree with therelationship of decomposition and AET found byMeentmeyer (1978). The effects of these variables are diffi-cult to separate, as they are not entirely independent. Astemperature decreases however, the contribution to the AET

function will certainly decline. Tundra sites are known tohave large variations in soil moisture and low mean dailytemperatures (Flanagan and Veum 1974). Given that thevariation in soil moisture exceeds the magnitude of tempera-ture variation at Niwot Ridge, the finding that moisture ef-fect exceeds that of temperature is not surprising. Moreover,a synthesis of data from circumpolar sites shows moisture tobe a primary factor controlling of litter decay rates in arcticsystems (Heal and French 1974). Furthermore, if moisturelimits microbial activity, temperature variation would havelittle effect other than to exacerbate the limitation throughincreased evaporation. This is supported by Flanagan and

© 1998 NRC Canada

Bryant et al. 1301

Fig. 4. Nitrogen dynamics of surface and buried fine-root litter in an alpine landscape. Native foliage (A) and buried fine-root(B) litters were incubated in three community types. LIDET litter (C, foliage; D, buried fine roots) were incubated in the mesicmeadow community. Error bars are SE.

Fig. 5. Relationship of initial nitrogen concentration onk factorsof individual replicates at 3 years for LIDET fine-root litter.

Fig. 6. Relationship of initial extractive fraction on net nitrogenimmobilization of LIDET foliage litter following 12 monthsincubation. Negative values denote N loss. Error bars are SE.

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

© 1998 NRC Canada

1302 Can. J. Bot. Vol. 76, 1998

Veum (1974) who found that the sensitivity of decomposi-tion to temperature increased with soil moisture in wetcoastal tundra.

Given the high concentration of extractive fractions in na-tive surface litter we cannot rule out leaching as a large por-tion of mass loss during this relatively short-term study.While microbial biomass does not imply microbial activity,the lack of a significant correlation of litter mass loss withmicrobial biomass and the late onset of N immobilizationsupports leaching as a major contributor to mass loss. Thestrong negative relationship of extractives with N immobili-zation (or rather N losses) is particularly indicative (Fig. 6).

This leaching loss of N may explain the observation thatN losses parallel mass loss as moisture availability in-creases. Fisk and Schimdt (1995) found a similar pattern ofN mineralization, suggesting that N cycling rates are linkedto moisture availability at this site.

Other nonmicrobial factors such as fragmentation mayalso be important, as evidenced by the differences in massloss among fine- and coarse-mesh litter bags. While thesource of this fragmentation is not known, communitionfrom microarthopods is a good possibility as the 0.3-mmmesh litter bags would exclude most prostigmatid, and prob-ably all orbatid, mites. Both are known to accelerate litterfragmentation (Seastedt and Crossley 1980) and have beencollected at this site (O’Lear and Seastedt 1994). Given theharsh environment of the alpine, other causes of fragmenta-tion such as wind, rain, hail, snow, or even trampling areequally likely. Fragmentation may perform a critical func-tion in alpine tundra by ameliorating conditions for decom-position. Aside from increased surface area, smaller particlesare more likely to be incorporated in the surface soil, wheremoisture availability for decomposer microbes would behigher than at the litter layer. Unfortunately, the litter bagtechnique does not measure decomposition dynamics at thisscale.

The observation that N additions do not increase massloss (Table 2) is consistent with earlier findings regardingexogenous N availability and litter decay rates(McClaugherty et al. 1985; Holland and Coleman 1987).The LIDET surface litters showed a weak (but not signifi-cant) correlation with litter N content and decomposition(Fig. 5). This evidence does not support the alternate hy-pothesis that N limits microbial decay of native litter. There-fore, the null hypothesis cannot be rejected. However, thedelay in nitrogen accumulation (Fig. 4) suggests that micro-bial growth on these litters may be retarded during the first 2years; thus, any effect of N availability may not yet be ap-parent. Future observations of the LIDET litters should pro-vide additional data to address the question of intrinsic Ncontent on decay and determine if additional long-term Naddition experiments are required.

While the AIS content of LIDET surface litters was foundto be the best predictor of decay (for the chemical compo-nents measured) (Fig. 4), the assumption that the recalcitrantfraction is hindering decay is premature at this stage. TheAIS fraction is inherently correlated with extractive andACS fractions, as any change in one fraction would necessi-tate an inverse change in one or both of the others. In addi-tion, the degree of mass lost after 3 years is still well withinthe labile carbon fraction of these litters. This relationship

could easily be expressed as a positive correlation ofk fac-tors with the sum of the labile pools (extractives + ACS or1 – AIS). The increase in the percent of original N since thesecond-year harvest (Fig. 4C) also suggests that the micro-bial community has not reached the point of relative C limi-tation often associated with N mineralization in recalcitrantresidues (Melillo and Aber 1984; McClaugherty and Berg1987; Aber et al. 1982, 1990).

Fine root decay followed a pattern similar to surface litterwith the exception of extremely slow rates (Figs. 1B and1D). Losses during the first year do not exceed the percentmass of the extractive fraction for any species (including na-tive fine roots). Curiously, no correlation was found for thisfraction and mass loss during any time period. However, Nlosses in the first year, the lack of correlation of mass losswith microbial counts, and the exclusion of microarthopodsby the fine-mesh root litter bags, suggest leaching as theprimary factor for mass loss in fine roots as well as surfacelitters.

The apparent negative effect of intrinsic N on root decay(Fig. 5) could be explained by inhibition of lignocellulolyticenzymes. Kirk (1987) found secretion of lignin–peroxidaseby certain species of white-rot fungi to be stimulated by Nstarvation and inhibited by the presence of N. While fungalbiomass counts on buried root litter were high at 15 months(Table 4), species composition was not determined, and thepresence of white rot fungi at this site is unknown. Further-more, it is not known how widespread this enzyme system isamong other species of fungi, nor the extent to which Navailability controls enzyme functions among fungal species(see Leatham and Kirk 1983). Given the previous evidencefor leaching losses, this observation could be explained bythe high extractive fraction inAndropogon gerardii. How-ever, no relationship of mass loss and extractive fraction wasfound for fine roots. Clearly, future investigations usingmore numerous and chemically varied species of root litterare required to test these hypotheses.

Decomposition of buried fine roots was markedly slowerthan any surface litter. Seastedt et al. (1992) observed thatroots decomposed at an equal or greater rate than surface fo-liage in grasslands regardless of the lower quality of rootsubstrate. As in grasslands, belowground biomass and pro-duction exceeds that for aboveground in arctic (Flanaganand Veum 1974) and alpine tundra (Webber and May 1977;Fisk 1995, Walker et al. 1994). The combination of higherbelowground production and lower decay rates suggest thatroot biomass may be a significant long-term reservoir of Cand N as well as the major contributor of these elements totundra soil reservoirs.

(1) Moisture availability is the primary controlling vari-able in the early stages of surface litter decomposition in al-pine tundra. This may be related to the large contribution ofleaching to mass losses in early stages of decay.

(2) Nitrogen additions do not have a measurable effect onnative litter decay at the Niwot Ridge site.

(3) Nitrogen release parallels mass loss across the alpinelandscape and compares well with landscape patterns of Ncycling found by other researchers (Fisk and Schmidt 1995).

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

N immobilization is delayed and thus does not stabilize Nlosses from litter as seen in other systems.

(4) Following an initial mass loss in the first growing sea-son, buried fine-root litter decay does not vary significantlyamong communities. Mass remaining appears to stabilize atapproximately 80% of original. While intrinsic N may influ-ence root decay, more research is required to discern possi-ble mechanisms involved.

(5) Surface foliage carbon turnover rates are greater thanthose of buried root litter. Taking into account greater pro-portional belowground production, and decreased decay ofroot litter, standing dead root biomass appears to be a long-term reservoir for C and may be a major contributor to soilorganic matter.

(6) The level of original N retained in decaying root litter(-50%) suggests that standing dead root biomass is also along-term reservoir for N in this system.

The authors thank Ed Rastetter and Bill Currie of theEcosystem Center, Woods Hole, Mass. and two anonymousreviewers for comments that greatly improved earlier ver-sions of the manuscript. John Aber and Steve Newman ofthe Complex Systems Research Center, University of NewHampshire, and Carol Wessman of The University of Colo-rado provided assistance and laboratory space for chemicalanalysis. Research was supported by the Niwot Ridge LongTerm Ecological Research program and the Long TermIntersite Decomposition Experiment Team (NSF grant No.BSR-9108329), both funded by the National Science Foun-dation.

Aber, J.D., and Melillo, J.M. 1982. Nitrogen immobilization indecaying hardwood leaf litter as a function of initial nitrogenand lignin content. Can. J. Bot.60: 2263–2269.

Aber, J.D., and Melillo, J.M. 1991. Terrestrial ecosystems.Saunders College Publishing, Philadelphia, Pa.

Aber, J.D., Melillo, J.M., and McClaugherty, C.A. 1990. Predictinglong-term patterns of mass loss, nitrogen dynamics and soil or-ganic matter formation from initial fine litter chemistry in tem-perate forest ecosystems. Can. J. Bot.68: 2201–2208.

Berendse, F., Berg, B., and Bosatta, E. 1987. The effect of ligninand nitrogen on decomposition of litter in nutrient-poor ecosys-tems: a theoretical approach. Can. J. Bot.65: 1116–1120.

Berg, B. 1987. Dynamics of N (15N) in decomposing Scots pineneedle litter. Long-term decomposition in a Scots pine forest.VI. Can. J. Bot.66: 1539–1546.

Bowman, W.D. 1992. Inputs and storage of N in winter snowpacksin an alpine ecosystem. Arct. Alp. Res.24: 211–215.

Effland, M.J. 1977. Modified procedure to determine acid insolu-ble lignin in wood and pulp. Tappi,60: 143–144.

Fisk, M.C. 1995. Nitrogen dynamics in an alpine tundra. Ph.D. dis-sertation, University of Colorado, Boulder, Colo.

Fisk, M.C., and Schmidt, S.K. 1995. Nitrogen turnover by sequen-tial immobilization and mineralization during residue decompo-sition in soils. Soil Sci. Soc. Am. J.59: 1036.

Flannagan, P.W., and Veum, A.K. 1974. Relationships between res-piration, weight loss, and temperature and moisture in organicresidues on tundra.In Soil organisms and decomposition in tun-

dra.Edited byA.J. Holding. Tundra Biome Steering Committee,Stockholm, Sweden.

Fog, K. 1988. The effect of added nitrogen on the rate of decompo-sition of organic matter. Biol. Rev. Cambridge Philos. Soc.66:433–462.

Fogel, R., and Cromack, K., Jr. 1977. Effect of habitat and sub-strate quality of Douglas-fir litter decomposition in western Ore-gon. Can. J. Bot.55:1632–1640.

Heal, O.W., and French, D.D. 1974. Decomposition of organicmatter in tundra.In Soil organisms and decomposition in tundra.Edited by A.J. Holding. Tundra Biome Steering Committee,Stockholm, Sweden.

Hobbie, S.E. 1996. Temperature and plant species control overlitter decomposition in Alaskan tundra. Ecol. Monogr.66:503–522.

Holland, E.A., and Coleman. D.C. 1987. Litter placement effectson microbial and organic matter dynamics in an agroecosystem.Ecology,68: 425–433.

Hunt, H.W., Ingham, E.R., Coleman, D.C., Elliott, E.T., and Reid,C.P.P. 1988. Nitrogen limitation of production and decomposi-tion in prairie mountain meadow and forest. Ecology,69:1009–1016.

Ingham, E.R., and Klein, D.A. 1984. Soil fungi: measurement ofhyphal lengths. Soil Biol. Biochem.12: 279–280.

Kelly, J.M., and Henderson, G.S. 1978. Effects of nitrogen andphosphorus additions on deciduous litter decomposition. SoilSci. Soc. Am J.42: 972–976.

Kirk, K.T. 1987. Enzymatic “combustion”: the microbial degrada-tion of lignin. Annu. Rev. Microbiol.41: 465–505.

Leathum, G.F., and Kirk, T.K. 1983. Regulation of lignolytic activ-ity by nutrient nitrogen in white-rot basidiomycetes. FEMSMicro. Lett. 16: 65–67.

Lewis, W.M., Grant, M.C., and Saunders, J.F., III 1984. Chemicalpatterns of bulk atmospheric deposition in the state of Colorado.Water Resour. Res.20: 1691–1704.

May, D.E., and Webber, P.J. 1975. Summary of soil and plantcanopy temperatures for the major vegetation types from NiwotRidge, CO, for the period July 1972 – October 1974. U.S.Tundra Biome Data Rep. No. 75-80.

McClaugherty, C.A., and Berg, B. 1987. Cellulose and nitrogenconcentrations as rate regulating factors in late stages of litterdecomposition. Pedobiologia,30: 101–112.

McClaugherty, C., Pastor, A.J., and Aber. J.D. 1985. Forest litterdecomposition in relation to soil nitrogen dynamics and litterquality. Ecology,66: 266–275.

Meentmeyer, V. 1978. Macroclimate and lignin control of litter de-composition rates. Ecology,59: 465–472.

Melillo, J.M., and Aber, J.D. 1984. Nutrient immobilization indecaying litter: an example of carbon–nutrient interactions.InTrends in ecological research in the 1980s.Edited byJ.H. Coleyand F.P. Goller. Plenum Press, New York. pp. 193–214.

Melillo, J.M., Aber J.D., and Muratore, J.F. 1982. Nitrogen andlignin control of hardwood leaf litter decomposition dynamics.Ecology,63: 621–626.

Melillo, J.M., Aber, J.D., Likens, A.E., Ricca, A., Fry, B., andNadelhoffer, K. 1989. Carbon and nitrogen dynamics along thedecay continuum: plant litter to soil organic matter.In Ecologyof arable lands.Edited by M. Clarholm and L. Bergstrom.Kluwer, London, U.K. p. 297.

Newman, S.D., Soulia, M.E., Aber, J.D., Dewey, B., and Ricca, A.1995. Analysis of forest foliage I: laboratory procedures forproximate carbon fraction and nitrogen determination. J. NearInfrared Spectrosc.21: 398–412.

© 1998 NRC Canada

Bryant et al. 1303

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.

© 1998 NRC Canada

1304 Can. J. Bot. Vol. 76, 1998

O’Lear, H.A., and Seastedt, T.R. 1994. Landscape patterns of litterdecomposition in alpine tundra. Oecologia,99: 95–99.

Schlesinger, W.H. 1991. Biogeochemistry, an analysis of globalchange. Academic Press, San Diego, Calif.

Seastedt, T.R., and Crossley, D.A., Jr. 1980. Effects ofmicroarthopods on the seasonal dynamics of nutrients in forestlitter. Soil Biol. Biochem.12: 333–342.

Seastedt, T.R., Parton, W.J., and Ojima, D.S. 1992. Mass loss andnitrogen dynamics of decaying litter of grasslands: the apparentlow nitrogen immobilization potential of root detritus. Can. J.Bot. 70: 384–391.

Sievering, H.D., Burton, D., and Caine, N. 1992. Atmosphericloading of nitrogen to alpine in the Colorado Front Range.Global Biogeochem. Cycles,6: 339–346.

Sokal, R.R., and Rohlf, F.S. 1995. Biometry, the principles andpractice of statistics in biological research. 2nd ed. W.H. Free-man, San Fransisco, Calif. 859 pp.

Swift, M.J., Heal O.W., and Anderson, J.M. 1979. Decompositionin terrestrial ecosystems. Stud. Ecol. No. 5.

Taylor, B., Parkinson, R.D., and Parsons, W.F. 1989. Nitrogen andlignin content as predictors of litter decay rates: a microcosmtest. Ecology,70: 97–104.

Taylor, R.V., and Seastedt, T.R. 1994. Short- and long-term pat-terns of soil moisture in alpine tundra. Arctic Alp. Res.26:14–20.

van Cleve, K. 1974. Organic matter quality in relation to decompo-sition. In Soil organisms and decomposition in tundra.Edited byA.J. Holding. Tundra Biome Steering Committee, Stockholm,Sweden.

van Vuuren, M.M.I., and van Der Veerden, L.J. 1992. Effects ofthree rates of atmospheric nitrogen deposition enriched with15Non litter decomposition in a heathland. Soil Biol. Biochem.24:527–532.

Vitousek, P.M., Turner, D.R., Parton, W.J., and Sanford, R.L. 1994.Litter decomposition on the Mauna Loa environmental matrix,Hawai’i: patterns, mechanisms and models. Ecology,75:418–429.

Walker, D.A., Halfpenny, J.C., Walker, M.D., and Wessman, C.A.1993. Long-term studies of snow vegetation interactions. Bio-Science,43: 287–301.

Walker, M.D., Webber, P.J., Arnold, E.H., and May, D.E. 1994. Ef-fects of interannual climate variations on phytomass in alpinetundra. Ecology,75: 393–408.

Webber, P.J., and Ebert May, D.C. 1977. The magnitude and distri-bution of belowground plant structures in the alpine tundra ofNiwot Ridge, Colorado. Arct. Alp. Res.9: 157–174.

Witkamp, M. 1966. Decomposition of leaf litter in relation to envi-ronment, microflora, and microbial respiration. Ecology,47:194–201.

Can

. J. B

ot. D

ownl

oade

d fr

om w

ww

.nrc

rese

arch

pres

s.co

m b

y U

nive

rsity

of

P.E

.I. o

n 11

/14/

14Fo

r pe

rson

al u

se o

nly.